© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 40976
    Prev
    Next
    INDEX
    Explanations

    references to justice and accountability in legal contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    èŃ°
    -0.16
    unes
    -0.14
    uling
    -0.14
     padded
    -0.14
    oyer
    -0.14
    igits
    -0.14
     resc
    -0.14
    ataka
    -0.13
    ecome
    -0.13
    ÑĢак
    -0.13
    POSITIVE LOGITS
     justice
    0.68
    justice
    0.52
     Justice
    0.52
    Justice
    0.46
     fairness
    0.38
     fair
    0.36
     injustice
    0.33
    fair
    0.32
     JUST
    0.29
     Fair
    0.28
    Activations Density 0.046%

    No Known Activations